Hi there community!
I am a student currently enrolled in a Research Traineeship Program. My lab studies Systemic Lupus Erythematosus (SLE), and as a part of this research we are investigating microbiome in SLE patients and comparing to control groups.
So far, we've done 16S rRNA amplicon sequencing of those patients. Some data were generated while I wasn't here (34 fecal samples, 15 controls, 19 SLE patients), and some are fairly recent (75 fecal samples, 25 controls, 50 SLE patients). Still not all samples have been sequenced, some failed PCR, and we'll be soon ordering mock communities in order to enhance our analyses, as some of the outcome results are pipeline-dependent (soft for OTU picking, filtering parameters etc.).
Since 16S rRNA seems to be inefficient in exploring this immune disease, and "predictive metagenomics" (with PICRUSt software) is far too inaccurate, since the sequences that fail to map to reference (GreenGenes) are thrown out, we've decided to go with Shotgun Sequencing.
Having read some papers (like: "How much metagenomic sequencing is enough to achieve a given goal?"), it is recommended to use at least 7Gbp (assuming x20 coverage to enumerate gene contents of prokaryotes with relative abundance of more than 1% in the human microbiota). Having 109 samples this is far too expensive. As a first step, I though about sticking to latest dataset, with 75 samples - i.e. reducing the number of samples to sequence. But this is too expensive (75 samples * 7GBp = 525).
I don't have that much of an experience here, therefore I'd like to ask you, dear community, to help me out here. Direct me to some materials, or provide some suggestions. I don't want to remove patients from my dataset, as statistic power would decrease, but at the same time reducing depth could lead to "shallow" data. What are you thoughts?
I have "paid" access to Illumina machines listed below:
Best
Robert.
I am a student currently enrolled in a Research Traineeship Program. My lab studies Systemic Lupus Erythematosus (SLE), and as a part of this research we are investigating microbiome in SLE patients and comparing to control groups.
So far, we've done 16S rRNA amplicon sequencing of those patients. Some data were generated while I wasn't here (34 fecal samples, 15 controls, 19 SLE patients), and some are fairly recent (75 fecal samples, 25 controls, 50 SLE patients). Still not all samples have been sequenced, some failed PCR, and we'll be soon ordering mock communities in order to enhance our analyses, as some of the outcome results are pipeline-dependent (soft for OTU picking, filtering parameters etc.).
Since 16S rRNA seems to be inefficient in exploring this immune disease, and "predictive metagenomics" (with PICRUSt software) is far too inaccurate, since the sequences that fail to map to reference (GreenGenes) are thrown out, we've decided to go with Shotgun Sequencing.
Having read some papers (like: "How much metagenomic sequencing is enough to achieve a given goal?"), it is recommended to use at least 7Gbp (assuming x20 coverage to enumerate gene contents of prokaryotes with relative abundance of more than 1% in the human microbiota). Having 109 samples this is far too expensive. As a first step, I though about sticking to latest dataset, with 75 samples - i.e. reducing the number of samples to sequence. But this is too expensive (75 samples * 7GBp = 525).
I don't have that much of an experience here, therefore I'd like to ask you, dear community, to help me out here. Direct me to some materials, or provide some suggestions. I don't want to remove patients from my dataset, as statistic power would decrease, but at the same time reducing depth could lead to "shallow" data. What are you thoughts?
I have "paid" access to Illumina machines listed below:
Best
Robert.